Abstract

The project subCULTron aims for achieving long-term autonomy in a learning, self-regulating, self-sustaining underwater society/culture of robots in a complex environment: the Lagoon of Venice. The heterogeneous system we are developing consists of 3 different agent types: aMussels, aFish and aPads. Artificial mussels (aMussels) are the collective long-term memory of the system, allowing information to stay beyond the runtime of other agents. These mussels monitor the natural habitat, including biological agents like algae, bacterial incrustation of the environment and fish. On the water surface, artificial lilypads (aPads) interface with the human society, delivering energy and information from the outside world. Between those two layers, artificial fish (aFish) move/monitor/explore the environment and exchange information with the mussels and lilypads.

1. Excellence

subCULTron aims for achieving long-term autonomy in a learning, self-regulating, self-sustaining underwater society/culture of robots in a high-impact application area: Venice, Italy.

Our heterogeneous system consists of 3 different agent types:

• On the seabed, artificial mussels are the collective long-term memory of the system, allowing information to stay beyond the runtime of other agents, thus allowing to continue learning from previously learned states. These mussels monitor the natural habitat, including biological agents like algae, bacterial incrustation and fish.
• On the water surface, artificial lily pads interface with the human society, delivering energy and information influx from ship traffic or satellite data.
• Between those two layers, artificial fish move/monitor/explore the environment and exchange info with the mussels and lily pads. Artificial mussels are a novel class of underwater agents.

The poject aimed to push forward the edge of knowledge with novel sensors (electric sense/electro-communication), novel bio-inspired algorithms (underwater hives) and novel energy harvesting in underwater scenarios. Our application field is a human- and animal-co-inhabited real-world environment of high impact: Venice canals & lagoon. These habitats are highly dynamic and structured, expected to be reflected by a spatial self-structuring of our mussel population. These sub-populations locally perform memetic or cultural learning algorithms on their specific local data. Thus our cultural evolution algorithms will promote sub-culture development, similar to the human society that does the same above the water level in parallel.

Overall, we aim for an artificial society underneath the water's surface working to benefit a human society above water.

2. Impact

A heterogeneous robotic swarm of over 130 agents was developed and tested within the scope of the project, with many novel algorithms applied to it.

The subCULTron project concerns itself with developing an underwater monitoring and exploration system and deploying it on an unsupervised mission of long-term marine monitoring and exploration. The subCULTron multi-agent system was envisioned as an artificial marine ecosystem. It is heterogeneous and consists of three agent types: five Autonomous Surface Vehicles (ASVs) called artificial lily pads (aPads), a small swarm of highly mobile artificial fish (aFish), and more than 100 underwater sensor nodes called artificial mussels (aMussels). aPads were designed as communication mediators between underwater robots and the surface and due to higher computational power as power and data hubs. Furthermore, they can relocate and charge surfaced underwater agents on demand. aMussels were designed as a low-cost long autonomy underwater sensor hub capable of vertical motion. Their goal is to continuously monitor the underwater environment while communicating collected data using short- or long-range communication interfaces. aFish is a robot similar to aMussel, with the capability of lateral movement at the expense of shorter autonomy. Their goal is to explore the environment outside the current deployment area. During exploration they also exchange necessary info with other agents in the swarm using short- or long-range communication interfaces.